mlflow API reference
400 public APIs from mlflow (mlflow/mlflow) — 118 classes, 77 functions, 205 methods. Signatures extracted by static analysis of the actual source.
Repository: mlflow/mlflow
| Kind | Count |
|---|---|
| Classes | 118 |
| Functions | 77 |
| Methods | 205 |
API list
func
.claude.skills.src.skills.commands.fetch_logs.to_seconds(ts:str) -> strTruncate timestamp to seconds precision for comparison.
method
.claude.skills.src.skills.github.client.GitHubClient.get_job(owner:str, repo:str, job_id:int) -> JobGet a specific job.
method
.claude.skills.src.skills.github.client.GitHubClient.get_job_run(owner:str, repo:str, run_id:int) -> JobRunGet a specific workflow run.
method
.claude.skills.src.skills.github.client.GitHubClient.get_jobs(owner:str, repo:str, run_id:int, attempt:int | None=None) -> AsyncIterator[Job]Get jobs for a workflow run.
method
.claude.skills.src.skills.github.client.GitHubClient.get_raw(endpoint:str) -> aiohttp.ClientResponseGet raw response for streaming.
method
.claude.skills.src.skills.github.client.GitHubClient.get_workflow_runs(owner:str, repo:str, head_sha:str | None=None, status:str | None=None) -> AsyncIterator[JobRun]Get workflow runs for a repository.
method
bin.install.Tool.get_extract_type(url:str) -> ExtractTypeInfer extract type from URL file extension.
method
bin.install.Tool.get_version_args() -> list[str]Get version check arguments, defaulting to --version.
func
dev.check_patch_prs.fetch_patch_prs(version:str) -> dict[int, bool]Fetch PRs labeled with `v{version}` from the MLflow repository.
func
dev.check_patch_prs.get_commits(branch:str) -> list[Commit]Get the commits in the release branch via GitHub API (last 90 days).
func
dev.check_patch_prs.validate_version(version:str) -> NoneValidate that the version has a micro version component.
class
dev.clint.src.clint.linter.PositionRepresents a position in source code with line and column.
class
dev.clint.src.clint.linter.RangeRepresents a range in source code with start and end positions.
method
dev.clint.src.clint.linter.Range.shift(offset:Position) -> 'Range'Shift this range by the given position offset.
method
dev.clint.src.clint.resolver.Resolver.clear() -> NoneClear all name mappings.
method
dev.clint.src.clint.resolver.Resolver.exit_scope() -> NoneExit current scope by restoring the previous snapshot.
method
dev.clint.src.clint.resolver.Resolver.resolve(node:ast.expr) -> list[str] | NoneResolve a node to its fully qualified name parts.
method
dev.clint.src.clint.resolver.Resolver.scope() -> Iterator[None]Context manager for automatic scope management.
method
dev.clint.src.clint.rules.lazy_import.LazyImport.check(module:str | None) -> boolCheck if importing the given module lazily should be flagged.
func
dev.clint.src.clint.utils.get_repo_root() -> PathFind the git repository root directory with caching.
func
dev.clint.src.clint.utils.resolve_expr(expr:ast.expr) -> list[str] | NoneResolves `expr` to a list of attribute names.
func
dev.detect_flaky_tests.attempt_jobs(repo:str, run_id:int, attempt:int) -> list[dict[str, Any]]All jobs for one run attempt (paginated: master has >20 jobs).
func
dev.detect_flaky_tests.failing_tests_from_log(repo:str, job_id:int) -> dict[str, str]{nodeid: first-line error} parsed from a job's log.
func
dev.detect_flaky_tests.gh_api_text_bytes(path:str) -> str | NoneFetch a possibly-binary endpoint (e.g.
func
dev.detect_flaky_tests.parse_failing_tests(log:str) -> dict[str, str]{nodeid: first-line error} parsed from raw Actions log text.
class
dev.dev_stubs.StubResultWhat a launcher must apply after installing stubs.
func
dev.dev_stubs.apply_to_environ(result:StubResult) -> NoneApply a StubResult's PATH prepend to ``os.environ`` in place.
func
dev.flavors.src.flavors._matrix.get_changed_flavors(changed_files:list[str], flavors:set[str]) -> set[str]Detects changed flavors from a list of changed files.
func
dev.flavors.src.flavors._matrix.get_latest_micro_versions(versions:list[Version]) -> list[Version]Returns the latest micro version in each minor version.
func
dev.flavors.src.flavors._update.update_version(src:str, key:str, new_version:str, category:str, update_max:bool) -> strExamples ======== >>> src = ''' ...
class
dev.gen_rest_api.APIMain API class for generating REST API documentation.
class
dev.gen_rest_api.MessageA protobuf message containing fields.
class
dev.gen_rest_api.ProtoEnumA protobuf enum with a series of Values.
class
dev.gen_rest_api.ServiceA protobuf service containing RPC methods.
class
dev.gen_rest_api.ValueAn enum value within a ProtoEnum.
class
dev.proto_plugin.ProtocPluginProtoc plugin implementation.
func
dev.pypi.src.pypi._client.get_package(name:str) -> PackageFetch package metadata from PyPI.
func
dev.remove_experimental_decorators.get_mlflow_release_dates() -> dict[str, datetime]Fetch MLflow release dates from PyPI API.
func
dev.remove_experimental_decorators.get_tracked_python_files() -> list[Path]Get all tracked Python files in the repository.
func
dev.remove_experimental_decorators.main() -> NoneMain entry point.
func
dev.update_changelog.batch_fetch_prs_graphql(pr_numbers:list[int]) -> list[PullRequest]Batch fetch PR data using GitHub GraphQL API.
func
dev.update_mlflow_versions.post_release(new_version:str) -> NoneUpdate MLflow package versions AFTER release.
func
dev.update_mlflow_versions.pre_release(new_version:str) -> NoneUpdate MLflow package versions BEFORE release.
func
dev.update_mlflow_versions.update_versions(new_py_version:str, helm_app_version:str | None=None) -> None`new_py_version` is either: - a release version (e.g.
func
fs2db.src.generate_synthetic_data.generate_logged_models(cfg:SizeConfig, experiments:list[ExperimentData]) -> list[str]Returns list of model artifact URIs.
func
fs2db.src.generate_synthetic_data.generate_traces(cfg:SizeConfig, experiments:list[ExperimentData]) -> list[str]Returns list of trace IDs.
class
mlflow.ag2.ag2_logger.ChatStateRepresents the state of a chat session.
func
mlflow.ai_commands.ai_command_utils.get_command(key:str) -> strGet command content by key.
func
mlflow.ai_commands.ai_command_utils.get_command_body(key:str) -> strGet command body content without frontmatter.
func
mlflow.ai_commands.ai_command_utils.list_commands(namespace:str | None=None) -> list[dict[str, Any]]List all available commands with metadata.
func
mlflow.ai_commands.ai_command_utils.parse_frontmatter(content:str) -> tuple[dict[str, Any], str]Parse frontmatter from markdown content.
func
mlflow.ai_commands.commands() -> NoneManage MLflow AI commands for LLMs.
func
mlflow.ai_commands.get_cmd(key:str) -> NoneGet a specific AI command by key.
func
mlflow.ai_commands.list_cmd(namespace:str | None) -> NoneList all available AI commands.
func
mlflow.ai_commands.run_cmd(key:str) -> NoneGet a command formatted for execution by an AI assistant.
func
mlflow.artifacts.list_artifacts(artifact_uri:str | None=None, run_id:str | None=None, artifact_path:str | None=None, tracking_uri:str | None=None) -> list[FileInfo]List artifacts at the specified URI.
func
mlflow.artifacts.load_dict(artifact_uri:str) -> dict[str, Any]Loads the artifact contents as a dictionary.
func
mlflow.artifacts.load_text(artifact_uri:str) -> strLoads the artifact contents as a string.
func
mlflow.assistant.clear_project_path_cache() -> NoneClear the project path cache to pick up config changes.
class
mlflow.assistant.cli.SpinnerSimple spinner animation for long-running operations.
func
mlflow.assistant.cli.commands(configure:bool)MLflow Assistant - AI-powered trace analysis.
class
mlflow.assistant.config.AssistantConfigMain configuration for MLflow Assistant.
method
mlflow.assistant.config.AssistantConfig.get_project_path(experiment_id:str) -> str | NoneGet the project path for a given experiment ID.
method
mlflow.assistant.config.AssistantConfig.get_selected_provider() -> ProviderConfig | NoneGet the currently selected provider.
method
mlflow.assistant.config.AssistantConfig.load() -> 'AssistantConfig'Load the assistant configuration from disk.
method
mlflow.assistant.config.AssistantConfig.save() -> NoneSave the assistant configuration to disk.
class
mlflow.assistant.config.PermissionsConfigPermission settings for the assistant provider.
class
mlflow.assistant.config.SkillsConfigSkills configuration for a provider.
func
mlflow.assistant.get_project_path(experiment_id:str) -> str | NoneGet the project path for a given experiment ID.
class
mlflow.assistant.providers.base.AssistantProviderAbstract base class for assistant providers.
method
mlflow.assistant.providers.base.AssistantProvider.description() -> strReturn a short description of the provider.
method
mlflow.assistant.providers.base.AssistantProvider.name() -> strReturn the provider identifier (e.g., 'claude_code').
method
mlflow.assistant.providers.base.AssistantProvider.resolve_skills_path(base_directory:Path) -> PathResolve the skills installation path.
class
mlflow.assistant.providers.base.CLINotInstalledErrorRaised when the provider CLI is not installed.
class
mlflow.assistant.providers.base.ProviderNotConfiguredErrorRaised when a provider is not properly configured.
func
mlflow.assistant.providers.base.clear_config_cache() -> NoneClear the config cache to pick up config changes.
class
mlflow.assistant.providers.claude_code.ClaudeCodeProviderAssistant provider using Claude Code CLI.
method
mlflow.assistant.providers.claude_code.ClaudeCodeProvider.check_connection(echo:Callable[[str], None] | None=None) -> NoneCheck if Claude CLI is installed and authenticated.
method
mlflow.assistant.providers.claude_code.ClaudeCodeProvider.resolve_skills_path(base_directory:Path) -> PathResolve the path to the skills directory.
func
mlflow.assistant.skill_installer.list_installed_skills(destination_path:Path) -> list[str]List installed skills in the specified destination path.
method
mlflow.assistant.types.Event.to_sse_event() -> strConvert the event to an SSE event string.
class
mlflow.assistant.types.MessageStructured message representation for assistant conversations.
class
mlflow.assistant.types.TextBlockText content block.
class
mlflow.assistant.types.ThinkingBlockThinking content block.
class
mlflow.assistant.types.ToolResultBlockTool result content block.
class
mlflow.assistant.types.ToolUseBlockTool use content block.
func
mlflow.claude_code.cli.commands()Commands for autologging with MLflow.
func
mlflow.claude_code.cli.stop_hook() -> NoneLegacy hook shim kept for older Python-hook installations.
class
mlflow.claude_code.config.TracingStatusDataclass for tracing status information.
func
mlflow.claude_code.config.get_env_var(var_name:str, default:str='') -> strGet environment variable with OS env taking highest priority.
func
mlflow.claude_code.config.get_tracing_status(settings_path:Path) -> TracingStatusGet current tracing status from Claude settings.
func
mlflow.claude_code.config.load_claude_config(settings_path:Path) -> dict[str, Any]Load existing Claude configuration from settings file.
func
mlflow.claude_code.config.save_claude_config(settings_path:Path, config:dict[str, Any]) -> NoneSave Claude configuration to settings file.
func
mlflow.claude_code.plugin.disable_tracing_plugin(settings_path:Path) -> boolRemove MLflow Claude config from settings.
func
mlflow.claude_code.tracing.get_logger() -> logging.LoggerGet the configured module logger.
func
mlflow.claude_code.tracing.setup_logging() -> logging.LoggerSet up logging directory and return configured logger.
func
mlflow.claude_code.tracing.setup_mlflow() -> NoneConfigure MLflow tracking URI and experiment.
func
mlflow.cli.crypto.commands()MLflow cryptographic management CLI.
func
mlflow.cli.datasets.commands()Manage GenAI evaluation datasets.
class
mlflow.cli.genai_eval_utils.AssessmentStructured assessment data for a trace evaluation.
class
mlflow.cli.genai_eval_utils.CellStructured cell data for table display with metadata.
class
mlflow.cli.genai_eval_utils.EvalResultContainer for evaluation results for a single trace.
class
mlflow.cli.genai_eval_utils.TableOutputContainer for formatted table data.
func
mlflow.cli.genai_eval_utils.format_table_output(output_data:list[EvalResult]) -> TableOutputFormat evaluation results as table data.
func
mlflow.cli.genai_eval_utils.resolve_scorers(scorer_names:list[str], experiment_id:str) -> list[Scorer]Resolve scorer names to scorer objects.
func
mlflow.cli.scorers.commands()Manage scorers, including LLM judges.
func
mlflow.cli.skills.view_command(skill_name:str)View the details of an MLflow skill.
func
mlflow.cli.traces.commands()Manage traces.
func
mlflow.cli.traces.delete_assessment(trace_id:str, assessment_id:str) -> NoneDelete an assessment from a trace.
func
mlflow.cli.traces.delete_trace_tag(trace_id:str, key:str) -> NoneDelete a tag from a trace.
func
mlflow.cli.traces.delete_traces(experiment_id:str, trace_ids:str | None=None, max_timestamp_millis:int | None=None, max_traces:int | None=None) -> NoneDelete traces from an experiment.
func
mlflow.cli.traces.get_assessment(trace_id:str, assessment_id:str) -> NoneGet assessment details as JSON.
func
mlflow.cli.traces.get_trace(trace_id:str, extract_fields:str | None=None, verbose:bool=False) -> NoneAll trace details will print to stdout as JSON format.
func
mlflow.cli.traces.set_trace_tag(trace_id:str, key:str, value:str) -> NoneSet a tag on a trace.
func
mlflow.cli.traces.update_assessment(trace_id:str, assessment_id:str, value:str | None=None, rationale:str | None=None, metadata:str | None=None) -> NoneUpdate an existing assessment.
method
mlflow.data.dataset.Dataset.digest() -> strA unique hash or fingerprint of the dataset, e.g.
method
mlflow.data.dataset.Dataset.name() -> strThe name of the dataset, e.g.
method
mlflow.data.dataset.Dataset.to_dict() -> dict[str, str]Create config dictionary for the dataset.
func
mlflow.data.dataset_registry.register_constructor(constructor_fn:Callable[[str | None, str | None], Dataset], constructor_name:str | None=None) -> strRegisters a dataset constructor.
method
mlflow.data.dataset_source.DatasetSource.load() -> AnyLoads files / objects referred to by the DatasetSource.
func
mlflow.data.dataset_source_registry.get_registered_sources() -> list[DatasetSource]Obtains the registered dataset sources.
func
mlflow.data.digest_utils.compute_numpy_digest(features, targets=None) -> strComputes a digest for the given numpy array.
func
mlflow.data.digest_utils.compute_pandas_digest(df) -> strComputes a digest for the given Pandas DataFrame.
func
mlflow.data.digest_utils.get_normalized_md5_digest(elements:list[Any]) -> strComputes a normalized digest for a list of hashable elements.
class
mlflow.data.evaluation_dataset.EvaluationDatasetAn input dataset for model evaluation.
method
mlflow.data.evaluation_dataset.EvaluationDataset.digest()Return the digest of the dataset.
method
mlflow.data.evaluation_dataset.EvaluationDataset.path()Dataset path
method
mlflow.data.evaluation_dataset.EvaluationDataset.predictions_name()return predictions name
method
mlflow.data.evaluation_dataset.EvaluationDataset.targets_name()return targets name
func
mlflow.data.get_source(dataset:DatasetEntity | DatasetInput | Dataset) -> DatasetSourceObtains the source of the specified dataset or dataset input.
class
mlflow.data.meta_dataset.MetaDatasetDataset that only contains metadata.
method
mlflow.data.meta_dataset.MetaDataset.schema() -> Any | NoneReturns the schema of the dataset.
method
mlflow.data.meta_dataset.MetaDataset.to_dict() -> dict[str, str]Create config dictionary for the MetaDataset.
class
mlflow.data.numpy_dataset.NumpyDatasetRepresents a NumPy dataset for use with MLflow Tracking.
method
mlflow.data.numpy_dataset.NumpyDataset.features() -> np.ndarray | dict[str, np.ndarray]The features of the dataset.
method
mlflow.data.numpy_dataset.NumpyDataset.profile() -> Any | NoneA profile of the dataset.
method
mlflow.data.numpy_dataset.NumpyDataset.source() -> DatasetSourceThe source of the dataset.
method
mlflow.data.numpy_dataset.NumpyDataset.targets() -> np.ndarray | dict[str, np.ndarray] | NoneThe targets of the dataset.
method
mlflow.data.numpy_dataset.NumpyDataset.to_dict() -> dict[str, str]Create config dictionary for the dataset.
class
mlflow.data.pandas_dataset.PandasDatasetRepresents a Pandas DataFrame for use with MLflow Tracking.
method
mlflow.data.pandas_dataset.PandasDataset.df() -> pd.DataFrameThe underlying pandas DataFrame.
method
mlflow.data.pandas_dataset.PandasDataset.predictions() -> str | NoneThe name of the predictions column.
method
mlflow.data.pandas_dataset.PandasDataset.profile() -> Any | NoneA profile of the dataset.
method
mlflow.data.pandas_dataset.PandasDataset.source() -> DatasetSourceThe source of the dataset.
method
mlflow.data.pandas_dataset.PandasDataset.targets() -> str | NoneThe name of the target column.
method
mlflow.data.pandas_dataset.PandasDataset.to_dict() -> dict[str, str]Create config dictionary for the dataset.
class
mlflow.data.polars_dataset.PolarsDatasetA polars DataFrame for use with MLflow Tracking.
method
mlflow.data.polars_dataset.PolarsDataset.df() -> pl.DataFrameUnderlying DataFrame.
method
mlflow.data.polars_dataset.PolarsDataset.predictions() -> str | NoneName of the predictions column.
method
mlflow.data.polars_dataset.PolarsDataset.profile() -> PolarsDatasetProfileProfile of the dataset.
method
mlflow.data.polars_dataset.PolarsDataset.source() -> DatasetSourceSource of the dataset.
method
mlflow.data.polars_dataset.PolarsDataset.targets() -> str | NoneName of the target column.
method
mlflow.data.polars_dataset.PolarsDataset.to_dict() -> PolarsDatasetConfigCreate config dictionary for the dataset.
class
mlflow.data.spark_dataset.SparkDatasetRepresents a Spark dataset (e.g.
method
mlflow.data.spark_dataset.SparkDataset.df()The Spark DataFrame instance.
method
mlflow.data.spark_dataset.SparkDataset.predictions() -> str | NoneThe name of the predictions column.
method
mlflow.data.spark_dataset.SparkDataset.profile() -> Any | NoneA profile of the dataset.
method
mlflow.data.spark_dataset.SparkDataset.schema() -> Schema | NoneThe MLflow ColSpec schema of the Spark dataset.
method
mlflow.data.spark_dataset.SparkDataset.source() -> SparkDatasetSource | DeltaDatasetSourceSpark dataset source information.
method
mlflow.data.spark_dataset.SparkDataset.to_dict() -> dict[str, str]Create config dictionary for the dataset.
class
mlflow.deployments.base.BaseDeploymentClientBase class exposing Python model deployment APIs.
method
mlflow.deployments.base.BaseDeploymentClient.list_deployments(endpoint=None)List deployments.
func
mlflow.deployments.base.target_help()..
func
mlflow.deployments.cli.commands()Deploy MLflow models to custom targets.
func
mlflow.deployments.cli.run_local(flavor, model_uri, target, name, config)Deploy the model locally.
class
mlflow.deployments.mlflow.MlflowDeploymentClientClient for interacting with the MLflow AI Gateway.
method
mlflow.deployments.mlflow.MlflowDeploymentClient.create_deployment(name, model_uri, flavor=None, config=None, endpoint=None)..
method
mlflow.deployments.mlflow.MlflowDeploymentClient.create_endpoint(name, config=None)..
method
mlflow.deployments.mlflow.MlflowDeploymentClient.delete_deployment(name, config=None, endpoint=None)..
method
mlflow.deployments.mlflow.MlflowDeploymentClient.delete_endpoint(endpoint)..
method
mlflow.deployments.mlflow.MlflowDeploymentClient.get_deployment(name, endpoint=None)..
method
mlflow.deployments.mlflow.MlflowDeploymentClient.list_deployments(endpoint=None)..
method
mlflow.deployments.mlflow.MlflowDeploymentClient.update_deployment(name, model_uri=None, flavor=None, config=None, endpoint=None)..
method
mlflow.deployments.mlflow.MlflowDeploymentClient.update_endpoint(endpoint, config=None)..
class
mlflow.deployments.openai.OpenAIDeploymentClientClient for interacting with OpenAI endpoints.
method
mlflow.deployments.openai.OpenAIDeploymentClient.create_deployment(name, model_uri, flavor=None, config=None, endpoint=None)..
method
mlflow.deployments.openai.OpenAIDeploymentClient.create_endpoint(name, config=None)..
method
mlflow.deployments.openai.OpenAIDeploymentClient.delete_deployment(name, config=None, endpoint=None)..
method
mlflow.deployments.openai.OpenAIDeploymentClient.delete_endpoint(endpoint)..
method
mlflow.deployments.openai.OpenAIDeploymentClient.get_deployment(name, endpoint=None)..
method
mlflow.deployments.openai.OpenAIDeploymentClient.list_deployments(endpoint=None)..
method
mlflow.deployments.openai.OpenAIDeploymentClient.update_deployment(name, model_uri=None, flavor=None, config=None, endpoint=None)..
method
mlflow.deployments.openai.OpenAIDeploymentClient.update_endpoint(endpoint, config=None)..
class
mlflow.diffusers.DiffusersAdapterModelA loaded LoRA adapter referencing a HuggingFace base model.
class
mlflow.dspy.callback.MlflowCallbackCallback for generating MLflow traces for DSPy components
func
mlflow.dspy.load.load_model(model_uri, dst_path=None)Load a Dspy model from a run.
func
mlflow.dspy.util.log_dspy_dataset(dataset:list['Example'], file_name:str)Log the DSPy dataset as a table.
func
mlflow.dspy.util.sanitize_params(params:dict[str, Any]) -> dict[str, Any]Sanitize the parameters by removing the sensitive parameters.
class
mlflow.dspy.wrapper.DspyChatModelWrapperMLflow PyFunc wrapper class for Dspy chat models.
class
mlflow.dspy.wrapper.DspyModelWrapperMLflow PyFunc wrapper class for Dspy models.
class
mlflow.entities._job.JobMLflow entity representing a Job.
method
mlflow.entities._job.Job.job_id() -> strString containing job ID.
method
mlflow.entities._job.Job.parsed_result() -> AnyReturn the parsed result.
method
mlflow.entities._job.Job.result() -> str | NoneString containing the job result or error message.
method
mlflow.entities._job.Job.retry_count() -> intInteger containing the job retry count
method
mlflow.entities._job.Job.timeout() -> float | NoneJob execution timeout in seconds.
method
mlflow.entities._job.Job.workspace() -> str | NoneWorkspace associated with this job.
class
mlflow.entities._job_status.JobStatusEnum for status of a Job.
method
mlflow.entities._job_status.JobStatus.from_int(status_int:int) -> 'JobStatus'Convert integer status to JobStatus enum.
method
mlflow.entities._job_status.JobStatus.from_str(status_str:str) -> 'JobStatus'Convert string status to JobStatus enum.
method
mlflow.entities._job_status.JobStatus.is_finalized(status:'JobStatus') -> boolDetermines whether or not a JobStatus is a finalized status.
method
mlflow.entities._job_status.JobStatus.to_int() -> intConvert JobStatus enum to integer.
method
mlflow.entities._job_status.JobStatus.to_proto() -> intConvert JobStatus enum to proto JobStatus enum value.
class
mlflow.entities.assessment.AssessmentBase class for assessments that can be attached to a trace.
class
mlflow.entities.assessment.ExpectationValueRepresents an expectation value.
class
mlflow.entities.assessment.FeedbackRepresents feedback about the output of an operation.
class
mlflow.entities.assessment.FeedbackValueRepresents a feedback value.
class
mlflow.entities.assessment.IssueReferenceRepresents a reference to an issue associated with a trace.
class
mlflow.entities.assessment.IssueReferenceValueRepresents an issue reference value.
class
mlflow.entities.assessment_source.AssessmentSourceSource of an assessment (human, LLM as a judge with GPT-4, etc).
class
mlflow.entities.assessment_source.AssessmentSourceTypeEnumeration and validator for assessment source types.
class
mlflow.entities.dataset.DatasetDataset object associated with an experiment.
method
mlflow.entities.dataset.Dataset.digest() -> strString digest of the dataset.
method
mlflow.entities.dataset.Dataset.name() -> strString name of the dataset.
method
mlflow.entities.dataset.Dataset.profile() -> strString profile of the dataset.
method
mlflow.entities.dataset.Dataset.schema() -> strString schema of the dataset.
method
mlflow.entities.dataset.Dataset.source() -> strString source of the dataset.
method
mlflow.entities.dataset.Dataset.source_type() -> strString source_type of the dataset.
class
mlflow.entities.dataset_input.DatasetInputDatasetInput object associated with an experiment.
method
mlflow.entities.dataset_input.DatasetInput.dataset() -> DatasetDataset.
method
mlflow.entities.dataset_input.DatasetInput.tags() -> list[InputTag]Array of input tags.
class
mlflow.entities.dataset_record.DatasetRecordRepresents a single record in an evaluation dataset.
class
mlflow.entities.dataset_record_source.DatasetRecordSourceSource of a dataset record.
class
mlflow.entities.dataset_record_source.DatasetRecordSourceTypeEnumeration for dataset record source types.
method
mlflow.entities.evaluation_dataset.EvaluationDataset.delete_records(record_ids:list[str]) -> intDelete specific records from the dataset.
method
mlflow.entities.evaluation_dataset.EvaluationDataset.from_dict(data:dict[str, Any]) -> 'EvaluationDataset'Create instance from dictionary representation.
method
mlflow.entities.evaluation_dataset.EvaluationDataset.from_proto(proto:ProtoDataset) -> 'EvaluationDataset'Create instance from protobuf representation.
method
mlflow.entities.evaluation_dataset.EvaluationDataset.merge_records(records:list[dict[str, Any]] | 'pd.DataFrame' | list['Trace']) -> 'EvaluationDataset'Merge new records with existing ones.
method
mlflow.entities.evaluation_dataset.EvaluationDataset.profile() -> str | NoneDataset profile information.
method
mlflow.entities.evaluation_dataset.EvaluationDataset.records() -> list[DatasetRecord]Get dataset records, loading them if necessary.
method
mlflow.entities.evaluation_dataset.EvaluationDataset.schema() -> str | NoneDataset schema information.
method
mlflow.entities.evaluation_dataset.EvaluationDataset.to_df() -> 'pd.DataFrame'Convert dataset records to a pandas DataFrame.
method
mlflow.entities.evaluation_dataset.EvaluationDataset.to_dict() -> dict[str, Any]Convert to dictionary representation.
method
mlflow.entities.evaluation_dataset.EvaluationDataset.to_proto() -> ProtoDatasetConvert to protobuf representation.
class
mlflow.entities.experiment.ExperimentExperiment object.
method
mlflow.entities.experiment.Experiment.experiment_id()String ID of the experiment.
method
mlflow.entities.experiment.Experiment.lifecycle_stage()Lifecycle stage of the experiment.
method
mlflow.entities.experiment.Experiment.name()String name of the experiment.
method
mlflow.entities.experiment.Experiment.trace_location() -> UnityCatalog | NoneTrace storage location, if configured.
method
mlflow.entities.experiment.Experiment.workspace() -> strWorkspace that owns the experiment, if known.
class
mlflow.entities.experiment_tag.ExperimentTagTag object associated with an experiment.
method
mlflow.entities.experiment_tag.ExperimentTag.key()String name of the tag.
method
mlflow.entities.experiment_tag.ExperimentTag.value()String value of the tag.
class
mlflow.entities.file_info.FileInfoMetadata about a file or directory.
method
mlflow.entities.file_info.FileInfo.file_size()Size of the file or directory.
method
mlflow.entities.file_info.FileInfo.path()String path of the file or directory.
class
mlflow.entities.gateway_budget_policy.BudgetActionAction to take when a budget is exceeded.
class
mlflow.entities.gateway_budget_policy.BudgetDurationUnitDuration unit for budget policy fixed windows.
class
mlflow.entities.gateway_budget_policy.BudgetTargetScopeTarget scope for a budget policy.
class
mlflow.entities.gateway_budget_policy.BudgetUnitBudget measurement unit.
class
mlflow.entities.gateway_budget_policy.GatewayBudgetPolicyRepresents a budget policy for the AI Gateway.
class
mlflow.entities.gateway_endpoint.FallbackConfigConfiguration for fallback routing strategy.
class
mlflow.entities.gateway_endpoint.FallbackStrategyFallback strategy for routing.
class
mlflow.entities.gateway_endpoint.GatewayEndpointModelConfigConfiguration for a model attached to an endpoint.
class
mlflow.entities.gateway_endpoint.GatewayModelDefinitionRepresents a reusable LLM model configuration.
class
mlflow.entities.gateway_endpoint.GatewayModelLinkageTypeType of linkage between endpoint and model definition.
class
mlflow.entities.gateway_endpoint.RoutingStrategyRouting strategy for gateway endpoints.
class
mlflow.entities.input_tag.InputTagInput tag object associated with a dataset.
method
mlflow.entities.input_tag.InputTag.key() -> strString name of the input tag.
method
mlflow.entities.input_tag.InputTag.value() -> strString value of the input tag.
method
mlflow.entities.issue.Issue.from_dictionary(issue_dict:dict[str, Any]) -> IssueCreate Issue from dictionary representation.
method
mlflow.entities.issue.Issue.from_proto(proto:ProtoIssue) -> IssueCreate Issue from protobuf representation.
method
mlflow.entities.issue.Issue.to_dictionary() -> dict[str, Any]Convert Issue to dictionary representation.
method
mlflow.entities.issue.Issue.to_proto() -> ProtoIssueConvert Issue to protobuf representation.
class
mlflow.entities.issue.IssueSeverityEnum for severity level of an :py:class:`mlflow.entities.Issue`.
class
mlflow.entities.issue.IssueStatusEnum for status of an :py:class:`mlflow.entities.Issue`.
class
mlflow.entities.logged_model_input.LoggedModelInputModelInput object associated with a Run.
method
mlflow.entities.logged_model_input.LoggedModelInput.model_id() -> strModel ID.
class
mlflow.entities.logged_model_output.LoggedModelOutputModelOutput object associated with a Run.
method
mlflow.entities.logged_model_output.LoggedModelOutput.model_id() -> strModel ID
method
mlflow.entities.logged_model_output.LoggedModelOutput.step() -> strStep at which the model was logged
class
mlflow.entities.logged_model_parameter.LoggedModelParameterMLflow entity representing a parameter of a Model.
method
mlflow.entities.logged_model_parameter.LoggedModelParameter.value()String value of the parameter.
class
mlflow.entities.logged_model_tag.LoggedModelTagTag object associated with a Model.
method
mlflow.entities.logged_model_tag.LoggedModelTag.key()String name of the tag.
method
mlflow.entities.logged_model_tag.LoggedModelTag.value()String value of the tag.
class
mlflow.entities.metric.MetricMetric object.
method
mlflow.entities.metric.Metric.dataset_digest() -> str | NoneString.
method
mlflow.entities.metric.Metric.dataset_name() -> str | NoneString.
method
mlflow.entities.metric.Metric.run_id() -> str | NoneString.
method
mlflow.entities.metric.Metric.step()Integer metric step (x-coordinate).
method
mlflow.entities.metric.Metric.value()Float value of the metric.
class
mlflow.entities.model_registry.model_version.ModelVersionMLflow entity for Model Version.
method
mlflow.entities.model_registry.model_version.ModelVersion.creation_timestamp() -> intInteger.
method
mlflow.entities.model_registry.model_version.ModelVersion.current_stage() -> str | NoneString.
method
mlflow.entities.model_registry.model_version.ModelVersion.description() -> str | NoneString.
method
mlflow.entities.model_registry.model_version.ModelVersion.last_updated_timestamp() -> int | NoneInteger.
method
mlflow.entities.model_registry.model_version.ModelVersion.model_id() -> str | NoneString.
method
mlflow.entities.model_registry.model_version.ModelVersion.name() -> strString.
method
mlflow.entities.model_registry.model_version.ModelVersion.run_id() -> str | NoneString.
method
mlflow.entities.model_registry.model_version.ModelVersion.run_link() -> str | NoneString.
method
mlflow.entities.model_registry.model_version.ModelVersion.source() -> str | NoneString.
method
mlflow.entities.model_registry.model_version.ModelVersion.status() -> strString.
method
mlflow.entities.model_registry.model_version.ModelVersion.status_message() -> str | NoneString.
method
mlflow.entities.model_registry.model_version.ModelVersion.user_id() -> str | NoneString.
method
mlflow.entities.model_registry.model_version.ModelVersion.version() -> strVersion
class
mlflow.entities.model_registry.model_version_tag.ModelVersionTagTag object associated with a model version.
method
mlflow.entities.model_registry.model_version_tag.ModelVersionTag.key()String name of the tag.
method
mlflow.entities.model_registry.model_version_tag.ModelVersionTag.value()String value of the tag.
class
mlflow.entities.model_registry.prompt.PromptEntity representing a prompt in the MLflow Model Registry.
method
mlflow.entities.model_registry.prompt.Prompt.creation_timestamp() -> int | NoneThe creation timestamp of the prompt.
method
mlflow.entities.model_registry.prompt.Prompt.description() -> str | NoneThe description of the prompt.
method
mlflow.entities.model_registry.prompt.Prompt.name() -> strThe name of the prompt.
method
mlflow.entities.model_registry.prompt.Prompt.tags() -> dict[str, str]Prompt-level metadata as key-value pairs.
class
mlflow.entities.model_registry.registered_model.RegisteredModelMLflow entity for Registered Model.
method
mlflow.entities.model_registry.registered_model.RegisteredModel.creation_timestamp()Integer.
method
mlflow.entities.model_registry.registered_model.RegisteredModel.description()String.
method
mlflow.entities.model_registry.registered_model.RegisteredModel.last_updated_timestamp()Integer.
method
mlflow.entities.model_registry.registered_model.RegisteredModel.name()String.
method
mlflow.entities.model_registry.registered_model.RegisteredModel.workspace() -> strWorkspace name for the registered model.
class
mlflow.entities.model_registry.registered_model_alias.RegisteredModelAliasAlias object associated with a registered model.
method
mlflow.entities.model_registry.registered_model_alias.RegisteredModelAlias.alias()String name of the alias.
class
mlflow.entities.model_registry.registered_model_tag.RegisteredModelTagTag object associated with a registered model.
method
mlflow.entities.model_registry.registered_model_tag.RegisteredModelTag.key()String name of the tag.
method
mlflow.entities.model_registry.registered_model_tag.RegisteredModelTag.value()String value of the tag.
class
mlflow.entities.param.ParamParameter object.
method
mlflow.entities.param.Param.value()String value of the parameter.
class
mlflow.entities.presigned_upload.CreatePresignedUploadResponseResponse from creating a presigned upload URL.
class
mlflow.entities.run.RunRun object.
method
mlflow.entities.run.Run.data() -> RunDataThe run data, including metrics, parameters, and tags.
method
mlflow.entities.run.Run.info() -> RunInfoThe run metadata, such as the run id, start time, and status.
method
mlflow.entities.run.Run.inputs() -> RunInputsThe run inputs, including dataset inputs.
method
mlflow.entities.run.Run.outputs() -> RunOutputsThe run outputs, including model outputs.
class
mlflow.entities.run_data.RunDataRun data (metrics and parameters).
class
mlflow.entities.run_info.RunInfoMetadata about a run.
method
mlflow.entities.run_info.RunInfo.artifact_uri()String root artifact URI of the run.
method
mlflow.entities.run_info.RunInfo.run_id()String containing run id.
method
mlflow.entities.run_info.RunInfo.run_name()String containing run name.
class
mlflow.entities.run_inputs.RunInputsRunInputs object.
method
mlflow.entities.run_inputs.RunInputs.dataset_inputs() -> list[DatasetInput]Array of dataset inputs.
method
mlflow.entities.run_inputs.RunInputs.model_inputs() -> list[LoggedModelInput]Array of model inputs.
class
mlflow.entities.run_outputs.RunOutputsRunOutputs object.
method
mlflow.entities.run_outputs.RunOutputs.model_outputs() -> list[LoggedModelOutput]Array of model outputs.
class
mlflow.entities.run_status.RunStatusEnum for status of an :py:class:`mlflow.entities.Run`.
class
mlflow.entities.run_tag.RunTagTag object associated with a run.
method
mlflow.entities.run_tag.RunTag.key()String name of the tag.
method
mlflow.entities.run_tag.RunTag.value()String value of the tag.
class
mlflow.entities.span.NoOpSpanNo-op implementation of the Span interface.
class
mlflow.entities.span.SpanA span object.
method
mlflow.entities.span.Span.attributes() -> dict[str, Any]Get all attributes of the span.
method
mlflow.entities.span.Span.end_time_ns() -> int | NoneThe end time of the span in nanosecond.
method
mlflow.entities.span.Span.events() -> list[SpanEvent]Get all events of the span.
method
mlflow.entities.span.Span.from_dict(data:dict[str, Any]) -> 'Span'Create a Span object from the given dictionary.
method
mlflow.entities.span.Span.from_dict_v2(data:dict[str, Any]) -> 'Span'Create a Span object from the given dictionary in v2 schema.
method
mlflow.entities.span.Span.get_attribute(key:str) -> Any | NoneGet a single attribute value from the span.
method
mlflow.entities.span.Span.inputs() -> AnyThe input values of the span.
method
mlflow.entities.span.Span.links() -> list['Link']Get all links of the span.
method
mlflow.entities.span.Span.llm_cost() -> dict[str, float] | NoneThe cost information for the span in USD.
method
mlflow.entities.span.Span.model_name() -> str | NoneThe model name used in the span.
method
mlflow.entities.span.Span.name() -> strThe name of the span.
method
mlflow.entities.span.Span.outputs() -> AnyThe output values of the span.
method
mlflow.entities.span.Span.parent_id() -> str | NoneThe span ID of the parent span.
method
mlflow.entities.span.Span.request_id() -> strDeprecated.
method
mlflow.entities.span.Span.span_id() -> strThe ID of the span.
method
mlflow.entities.span.Span.span_type() -> strThe type of the span.
method
mlflow.entities.span.Span.start_time_ns() -> intThe start time of the span in nanosecond.
method
mlflow.entities.span.Span.status() -> SpanStatusThe status of the span.
class
mlflow.entities.span.SpanTypePredefined set of span types.
func
mlflow.entities.span.create_mlflow_span(otel_span:Any, trace_id:str, span_type:str | None=None) -> Union['Span', 'LiveSpan', 'NoOpSpan']Factory function to create a span object.
class
mlflow.entities.span_event.CustomEncoderCustom encoder to handle json serialization.
class
mlflow.entities.span_log_level.SpanLogLevelLog level (severity) for an MLflow trace span.
class
mlflow.entities.span_status.SpanStatusStatus of the span or the trace.
method
mlflow.entities.span_status.SpanStatus.from_otel_status(otel_status:trace_api.Status) -> SpanStatusConvert OpenTelemetry status object to our status object.
class
mlflow.entities.span_status.SpanStatusCodeEnum for status code of a span
class
mlflow.entities.trace.TraceA trace object.
class
mlflow.entities.trace_data.TraceDataA container object that holds the spans data of a trace.
method
mlflow.entities.trace_data.TraceData.intermediate_outputs() -> dict[str, Any] | None..
method
mlflow.entities.trace_info.TraceInfo.cost() -> dict[str, float] | NoneReturns the aggregated cost for the trace in USD.
method
mlflow.entities.trace_info.TraceInfo.from_dict(d:dict[str, Any]) -> 'TraceInfo'Create a TraceInfoV3 object from a dictionary.
method
mlflow.entities.trace_info.TraceInfo.request_id() -> strDeprecated.
method
mlflow.entities.trace_info.TraceInfo.request_metadata() -> dict[str, str]Deprecated.
method
mlflow.entities.trace_info.TraceInfo.status() -> TraceStatusDeprecated.
method
mlflow.entities.trace_info.TraceInfo.to_dict() -> dict[str, Any]Convert the TraceInfoV3 object to a dictionary.
method
mlflow.entities.trace_info.TraceInfo.token_usage() -> dict[str, int] | NoneReturns the aggregated token usage for the trace.
class
mlflow.entities.trace_info_v2.TraceInfoV2Metadata about a trace.
method
mlflow.entities.trace_info_v2.TraceInfoV2.trace_id() -> strReturns the trace ID of the trace info.
class
mlflow.entities.trace_location.MlflowExperimentLocationRepresents the location of an MLflow experiment.
class
mlflow.entities.trace_location.TraceLocationRepresents the location where the trace is stored.
class
mlflow.entities.trace_location.TraceLocationBaseBase class for trace location classes.
class
mlflow.entities.trace_state.TraceStateEnum representing the state of a trace.
class
mlflow.entities.trace_status.TraceStatusEnum for status of an :py:class:`mlflow.entities.TraceInfo`.
class
mlflow.entities.view_type.ViewTypeEnum to filter requested experiment types.
class
mlflow.entities.webhook.WebhookMLflow entity for Webhook.
class
mlflow.entities.webhook.WebhookEventRepresents a webhook event with a resource and action.
class
mlflow.entities.webhook.WebhookTestResultMLflow entity for WebhookTestResult.
class
mlflow.entities.workspace.TraceArchivalConfigPython-facing configuration for workspace trace archival.
class
mlflow.entities.workspace.WorkspaceMinimal metadata describing a workspace.
class
mlflow.error_classification.ErrorClassError class names for MLflow error classification.
class
mlflow.error_classification.SqlStateSQLSTATE codes for MLflow error classification.
class
mlflow.evaluation.assessment.AssessmentAssessment data associated with an evaluation result.
method
mlflow.evaluation.assessment.Assessment.error_code() -> str | NoneThe error code.
method
mlflow.evaluation.assessment.Assessment.error_message() -> str | NoneThe error message.
method
mlflow.evaluation.assessment.Assessment.from_dictionary(assessment_dict:dict[str, Any]) -> 'Assessment'Create an Assessment object from a dictionary.
method
mlflow.evaluation.assessment.Assessment.metadata() -> dict[str, Any]The metadata associated with the assessment.
method
mlflow.evaluation.assessment.Assessment.name() -> strThe name of the assessment.
method
mlflow.evaluation.assessment.Assessment.rationale() -> str | NoneThe rationale / justification for the assessment.
method
mlflow.evaluation.assessment.Assessment.source() -> AssessmentSourceThe source of the assessment.
method
mlflow.evaluation.assessment.Assessment.value() -> bool | float | strThe assessment value.
class
mlflow.evaluation.assessment.AssessmentEntityAssessment data associated with an evaluation.
method
mlflow.evaluation.assessment.AssessmentEntity.boolean_value() -> bool | NoneThe boolean assessment value.
About this data
These signatures were extracted from the public source of mlflow/mlflow
using Python's ast module. Argument names, default values,
type annotations and return types are taken verbatim from the code.
Implementation bodies are never stored. See
how it works for details.